Architecture matters because while deep learning can conceivably fit a curve with a single, huge layer (in theory... Universal approximation theorem), the amount of compute and data needed to get there is prohibitive. Having a good architecture means the theoretical possibility of deep learning finding the right N dimensional curve becomes a practical reality.
Another thing about the architecture is we inherently bias it with the way we structure the data. For instance, take a dataset of (car) traffic patterns. If you only track the date as a feature, you miss that some events follow not just the day-of-year pattern but also holiday patterns. You could learn this with deep learning with enough data, but if we bake it into the dataset, you can build a model on it _much_ simpler and faster.
So, architecture matters. Data/feature representation matters.
But not all things you might do with a dotfile (or, more generally, per-user customization) are just replacing files. Things like cronjobs, brew installs, `defaults` in MacOS, etc. Viewing dotfile-based customization as strictly files to obliterate with pre-existing files is needlessly myopic.
For this broader problem, there are other more complete solutions that are more robust and flexible. Personally I like dotbot (https://github.com/anishathalye/dotbot) as a balance between power and simplicity, particularly when managing files across multiple OS homedirs (e.g. linux server, macos laptop).
Another thing about the architecture is we inherently bias it with the way we structure the data. For instance, take a dataset of (car) traffic patterns. If you only track the date as a feature, you miss that some events follow not just the day-of-year pattern but also holiday patterns. You could learn this with deep learning with enough data, but if we bake it into the dataset, you can build a model on it _much_ simpler and faster.
So, architecture matters. Data/feature representation matters.